Rail Transit In America -- A Comprehensive Evaluation of Benefits

This graph compares different categories of cities by various performance indicators. The dashed line at 100% indicates "Bus Only" city values.

These benefits cannot be attributed entirely rail transit. They partly reflect the larger average size of Large Rail cities. But taking size into account, cities with large, well-established rail transit systems still perform better in various ways than cities that lack rail systems. These benefits result from rail’s ability to help create more accessible land use patterns and more diverse transport systems.

Although Large Rail cities have higher per capita congestion costs, this occurs because congestion tends to increase with city size. Taking city size into account, rail transit turns out to significantly reduce per capita congestion costs, as indicated in Figure ES-3. Matched pair analysis indicates that Large Rail cities have about half the per capita congestion costs as other comparable size cities.

U.S. rail transit services require about $12.5 billion annual public subsidy (total capital and operating expenses minus fares), about an extra $90 per Large Rail city resident. However, economic benefits more than repay these subsidies: rail transit services are estimated to provide $19.4 billion in annual congestion cost savings, $8.0 billion in roadway cost savings, $12.1 billion in parking cost savings, $22.6 billion in consumer cost savings, and $5.6 billion in traffic accident cost savings. Rail transit also tends to provide economic development benefits, increasing business activity and tax revenues. It can be a catalyst for community redevelopment. Additional, potentially large benefits include improved mobility for non-drivers, increased community livability and improved public health.

This study critiques studies which imply that rail transit is ineffective. It finds that their analysis is often incomplete, inaccurate, and biased. It examines various factors that could offset rail transit benefits, including the possibility that transit oriented development is harmful to consumers, that new rail systems cannot achieve significant benefits, that apparent benefits of rail actually reflect other factors such as city size, and that bus transit can provide equal benefits at less cost.

This study indicates that rail transit is particularly important in large, growing cities. Large cities with well established rail systems are clearly advantaged in terms of congestion costs, consumer costs and traffic crash rates compared with cities that lack such systems. Cities with newer and smaller systems have not yet achieved the full impacts, but, if their rail systems continue to develop, their benefits should increase for decades, and so are a valuable legacy for the future.

This analysis does not mean that every rail transit project is cost-effective, or that rail is always better than bus or highway improvements. It attempts to provide a fair and balanced evaluation of the advantages and disadvantages of each mode, and identify situations in which each is most appropriate. This study concludes that rail transit provides significant benefits, particularly if implemented with supportive transport and land use policies. In many situations, rail transit is the most cost effective way to improve urban transportation.

During the last century most North American cities became increasingly automobile oriented (for this analysis "automobile" refers to any personal motor vehicle, including cars, light trucks, vans, SUVs and even motorcycles). Now, the majority of personal travel is by automobile, the majority of transportation resources (money and land) are devoted to automobiles and their facilities, and many communities have dispersed land use patterns that depend on automobile travel for access. The resulting growth in vehicle traffic creates various problems, including congestion, high road and parking facility costs, costs to consumers of owning and operating automobiles, traffic accidents, inadequate mobility for non-drivers, and various environmental impacts.

In recent years many experts and citizens have advocated diversifying our transport systems by increasing support for alternatives modes such as walking, cycling and public transit. To accomplish this many cities are making significant investments in public transit, including busways, light rail and heavy rail systems. There is considerable debate over the merits of these investments. Critics argue they are inappropriate and wasteful.

This study evaluates rail transit benefits based on a comprehensive analysis of transportation system performance in U.S. cities. It uses best available evaluation methods, based on guidance from leading experts and organizations (Cambridge Systematics, 1998; FTA, 1998: Lewis and Williams, 1999; Phillips, Karachepone and Landis, 2001; HLB, 2002; Kittleson & Associates, 2003; MKI, 2003; Litman, 2004a). This analysis takes into account a variety of performance factors, including the amount and type of travel that occurs, congestion costs, road and parking facility costs, consumer costs, accident rates, transit system efficiency and cost recovery, and various other impacts.

This study compares rail and bus transit, identifies the conditions in which each is most appropriate, and discusses the role that each mode can play in an efficient transportation system. It also describes various ways of improving transit service performance in order to increase benefits.

This study evaluates various criticisms of rail transit, including claims that it provides minimal congestion and emission reduction benefits, that it is not cost effective, and that money is better spent on roads, bus service or subsidized cars. It also examines various factors that could offset rail transit benefits, including the possibility that transit oriented development is harmful to consumers, that new rail systems cannot achieve significant benefits, that apparent benefits of rail actually reflect other factors such as city size, and that bus transit can provide equal benefits at less cost.

This section describes the evaluation of rail transit benefits. For more information on methodologies see, "Evaluating Public Transit Benefits and Costs" (www.vtpi.org/tranben.pdf). Analysis data are available in the "Transit Evaluation Spreadsheet" (www.vtpi.org/transit.xls).

About two dozen U.S. cities have some sort of rail transit service, but most are small and so cannot be expected to significantly effect regional transportation system performance, although they may have significant impacts on a particularly corridor or within a particular area. For this study, U.S. cities are divided into three categories:

0. Large Rail – Rail transit is a major component of the transportation system.

1. Small Rail – Rail transit is a minor component of the transportation system.

2. Bus Only – City has no rail transit system.

Seven cities are classified as "Large Rail," meaning that transit represents more than 20% of total commutes, and more than half of transit passenger-miles are by rail, as illustrated in Figure 1.

Figure 1 Transit Commute Mode Share (FTA, 2001)

This figure shows the portion of commutes by rail and bus transit. Only a few cities have rail systems large enough to significantly impact regional transportation system performance.

The next section of this report evaluates these different categories in terms of various transportation system performance indicators. Because Large Rail cities are relatively large, most comparisons include just the 50 largest cities to avoid skewing results with numerous small cities.

A key issue in evaluating transit is the degree to which it attracts riders and substitutes for automobile travel, and therefore reduces traffic problems such as congestion, parking costs and accidents. Rail tends to provide higher quality service than bus transit. Rail is usually more comfortable, faster (particularly if grade separated, so trains are not delayed by congestion) and better integrated into the urban landscape. As a result, rail transit usually attracts more riders within a given area, particularly discretionary riders (travelers who could drive but choose to ride transit, also called choice riders), and so is more effective than bus transit at reducing automobile trips (Pratt, 1999; FTA, 2002).

One recent study found that a 10% increase in a city’s rail transit service reduces 40 annual vehicle miles of travel per capita (70 VMT if New York City is included in the analysis), compared with just a one mile reduction from a 10% increase in bus service (Bento, et al, 2004). According to the Transit Performance Monitoring System (FTA, 2002), more than half of transit passengers report that without transit they would travel by automobile, either as a driver or passenger (some passenger trips would be ridesharing, using an otherwise empty seat without increasing vehicle mileage, while others would be chauffeured trips that do increase vehicle travel). Below is what respondents report they would do if transit service were unavailable, for all transit systems surveyed. Automobile substitution rates are higher in larger cities.

Other studies find similar results. A user survey in Vancouver, Canada found that 42% of Skytrain (rail) riders would otherwise drive, compared with 25-35% of bus riders. The table below provides information on the mode shifts that result from improved bus and rail transit service. These studies suggests that more than half of rail transit trips substitute for an automobile trip.

Table 1 Mode Shifts By New Transit Users (Pratt, 1999, Table 9-10)

Rail transit tends to leverage additional automobile travel reductions by providing a catalyst for more accessible land use patterns and reduced per capita vehicle ownership. This reflects the impacts of Transit Oriented Development (also called New Urbanism and Smart Growth), which consists of compact, walkable, mixed-use centers (TCRP, 2004). If you live near a rail transit station your neighborhood probably has a variety of shops and services nearby, and pedestrian-friendly streets, so you are more likely to walk for errands such as picking up a video or taking children to school, and your household may own fewer cars than it in a more automobile-dependent location. There are many types of transit oriented development, ranging from high-density commercial centers to small suburban villages. Many older urban neighborhoods that developed along streetcar lines retain transit oriented features decades after the rail transit service discontinued. Many of these are considered desirable neighborhoods due to those features.

Travel surveys find that households located near rail transit stations tend to own fewer cars and drive less than households in areas that lack rail service. This could partly reflect self-selection (households that prefer transit choose to live in such areas), but there is evidence that residents often reduce their vehicle ownership and shift travel patterns when they move there. A study of Orenco Station, a transit oriented development on Portland’s light rail line, found that 22% of residents commute by public transit, far higher than the 5% regional average, and 69% use public transit more often than they did in their previous community (Podobnik, 2002). In addition, the probability of a household owning a motor vehicle decreases by about a third for residents of such communities, taking into account other demographic and economic factors (Hess and Ong, 2002).

In other words, rail transit reduces automobile travel in two different ways: directly, when a rail passenger-mile substitutes for an automobile vehicle-mile, and indirectly when it creates more accessible land use and reduces automobile ownership in an area. Although indirect effects are difficult to measure, this and other studies suggest that they are often larger than direct effects. Research indicates that each rail transit passenger-mile represents a reduction of 3 to 6 automobile vehicle-miles, as summarized in Table 2, and in studies by Neff (1996) and Newman and Kenworthy (1999, p. 87).

Table 2 Transit Leverage: VMT Reductions Due to Transit (Holtzclaw 2000)

A key question is whether new rail systems significantly affect transportation and land use patterns within an acceptable time period, since land use patterns generally change slowly. Evidence from some cities indicates that they can. As described above, Portland has several new transit oriented neighborhoods where residents tend to own fewer cars and drive less, and rail ridership there is growing steadily, as shown in Figures 2.

Figure 27 Portland Transit Ridership Trends (APTA Data)

Bus transit does not generally effect land use in this way, and so does not seem to have a leverage effect on vehicle miles traveled. It is possible that bus transit programs that include incentives such as parking cash-out and location-efficient development (VTPI, 2004) could reduce vehicle travel and change land use similar to rail, but these impacts would result from the incentives, not the bus service itself.

Figure 3 Per Capita Transit Travel (FTA, 2001)

This analysis finds that per-capita transit ridership is far higher in rail transit cities, as illustrated in Figures 3 and 4. Annual per capita transit passenger-miles average 589 in Large Rail cities (520 excluding New York), 176 passenger-miles in Small Rail cities, and 118 passenger-miles in Bus Only cities. Although this partly reflects the tendency of transit ridership to increase with city size, cities with rail systems tend to occupy the upper-left area of the graph in Figure 3, indicating high ridership for their population.

Figure 4 Annual Per Capita Transit Ridership

Figure 5 Transit Commute Share (Census, 2002)

Figure 5 shows transit commute mode share for the 50 largest U.S. cities, indicating much higher rates for Large Rail cities. Large Rail cities have 34.8% transit mode share (30.7% excluding New York), as opposed to 11.0% for Small Rail and 4.5% for Bus Only cities. Although this can be partly explained by differences in city size, the graph shows that Large Rail cities tend to use transit far more than residents of comparable size cities that lack such systems. Transit mode share tends to be even higher for peak-period travel on rail transit corridors and destinations, such as downtowns.

Figure 6 Transit Commute Mode Share

Figure 7 Per Capita Vehicle Ownership (BLS, 2003)

Figure 7 shows how per capita vehicle ownership declines with rail transit. In Large Rail cities residents own 0.68 vehicles per capita (0.71 excluding New York), as opposed to 0.77 in Small Rail cities, and 0.80 in Bus Only cities. This is particularly notable because Large Rail city residents have higher average incomes than residents of other types of cities, which generally increases vehicle ownership. This reduction in vehicle ownership provides consumer cost savings and helps leverage additional reductions in automobile travel beyond just the passenger-miles shifted from driving to transit, as discussed elsewhere in this report.

Figure 8 Per Capita Private Vehicle Ownership

Figure 9 Average Per Capita Annual Vehicle Mileage (FHWA, 2002, Table 71)

Figure 9 shows average annual per capita vehicle mileage for various cities. Residents of Large Rail cities drive an average of 7,548 vehicle-miles (7,840 excluding New York), residents of Small Rail cities average 8,679 vehicle-miles, and residents of Bus Only cities average 9,506 annual vehicle-miles. Large Rail city residents drive 12% less per year than residents of Small Rail cities, and 20% less than residents of Bus Only cities. This indicates the leverage effect of rail. Residents of Large Rail cities average 470 more transit passenger-miles than Bus Only cities, and drive 1,958 fewer vehicle-miles, a 4:1 ratio. This ratio increases to 5:1 when the analysis is limited to cities with more than 2 million population, indicating that city size by itself does not explain these differences.

Figure 10 Annual Per Capita Vehicle-Miles

Traffic congestion consists of the incremental delay, stress, vehicle operating costs and pollution that each additional vehicle imposes on other road users. Congestion reduction is a primary transportation improvement objective. Special care is needed to accurately evaluate transit congestion reduction impacts ("Congestion Costs," Litman, 2003b). Traffic congestion tends to increase with city size because there are more vehicles within a given area. Rail transit systems are generally developed as cities grow large enough to experience significant congestion problems, so cities with rail transit tend to have worse congestion than those without, but it is wrong to suggest that rail transit causes congestion, or that congestion problems would be as bad if rail transit did not exist.

Congestion is a non-linear function: once a roadway reaches capacity even a small reduction in volumes can significantly reduce delays. For example, a 5% reduction in peak-hour traffic volumes on a road at 90% capacity can reduce delay by 20% or more. Transit can provide significant congestion reduction benefits, even if it only carries a small portion of total regional travel, because it offers an alternative on the most congested corridors. Reducing just a few percent of vehicles on such roads can significantly reduce congestion costs.

Congestion reduction benefits can be difficult to evaluate because urban traffic tends to maintain equilibrium: traffic volumes grow until congestion delays discourage additional peak-period trips. Grade-separated transit acts as a pressure-relief value, reducing the point of congestion equilibrium, as described in the box below. Although congestion never disappears, it is far less intense than would occur if such transit did not exist.

How Transit Reduces Traffic Congestion Urban traffic congestion tends to maintain equilibrium. If congestion increases, people change destinations, routes, travel time and modes to avoid delays, and if it declines they take additional peak-period trips. If roadway capacity increases, it will be partly filled by this latent demand (potential additional peak-period vehicle trips). Reducing this point of equilibrium is the only way to reduce congestion over the long run. The quality of travel alternatives has a significant effect on this equilibrium: If alternatives are inferior, few motorists will shift mode and the level of equilibrium will be high. If travel alternatives are relatively attractive, more motorists will shift modes, resulting in a lower equilibrium. Improving travel options can therefore benefit all travelers on a corridor, both those who shift modes and those who continue to drive. Shifts to alternative modes not only reduce congestion on a particular highway, they also reduce traffic discharged onto surface streets, providing "downstream" congestion reduction benefits. To reduce congestion, transit must attract discretionary riders (travelers who have the option of driving), which requires fast, comfortable, convenient and affordable service. When transit is faster and more comfortable than driving, a portion of travelers shift mode until congestion declines to the point that transit is no longer faster. As a result, the faster and more comfortable the transit service, the faster the traffic speeds on parallel highways. This theory is supported by studies which find that door-to-door travel times for motorists tend to converge with those of grade-separated transit (Mogridge, 1990; Lewis and Williams, 1999), and by studies such as this one, which find that congestion costs decline in cities with grade-separated transit systems.

The Texas Transportation Institute’s annual Urban Mobility Study (TTI, 2003) is the most commonly-used reference for comparing congestion costs between U.S. cities. It provides seven congestion indicators. Some of these indicators are more appropriate than others for evaluating transit impacts. Per-capita Congestion Cost is a better indicator of transit congestion reduction benefits, since it accounts for time savings that result from shifts to alternative modes and more accessible land use patterns. Measured in this way, Large Rail cities have substantially less congestion than other comparable size cities, as illustrated in Figure 11. For cities with Small Rail or Bus Only transit systems, traffic congestion increases substantially with city size, but cities with Large Rail transit systems do not follow this pattern.

Figure 11 Congestion Costs (TTI, 2003)

Detailed analysis of TTI data by Winston and Langer (2004) also indicates that both motorist and truck congestion costs decline in a city as rail transit mileage expands, but congestion costs increase with as bus transit mileage expands. This appears to occur because buses attract fewer travelers from driving, contribute to traffic congestion themselves, and have less positive impact on land use accessibility. Garrett (2004) found that traffic congestion growth rates declined somewhat in some U.S. cities after light rail service began. In Baltimore the congestion index increased an average of 2.8% annually before light rail, but only 1.5% annually after. In Sacramento the index grew 4.5% annually before light rail, but only 2.2% after. In St. Louis the index grew an average of 0.89% before light rail, and 0.86% after. In Dallas, the growth rate did not change.

Figure 12 Transit Congestion Cost Savings (TTI, 2003)

TTI estimates congestion cost savings from public transit services. Figures 12 and 13 compare this benefit for various cities. Large Rail cities have much greater transit congestion reductions than other cities. Of the 50 largest cities, Large Rail cities average $279 savings per capita, compared with $88 Small Rail cities, and $41 for Bus Only cities. These savings total more than $14.0 billion in Large Rail cities, $5.4 billion in Small Rail cities, and $1.8 billion dollars in Bus Only cities (considering only the 50 largest U.S. cities), indicating that rail provides $19.4 billion annual congestion cost savings. These savings approximately equal total U.S. public transit subsidies.

Figure 13 Transit Congestion Cost Savings

Matched pair analysis is used to determine whether these differences in congestion costs result from differences in city size. Three Large Rail cities (New York, Chicago and Philadelphia) are compared individually with three similar size Small Rail cities (Los Angeles, Miami and Dallas). The three Small Rail cities experience about twice the congestion delays as their matched Large Rail cities.

Table 3 Congestion Delay In Six Largest U.S. Cities

This leaves little doubt that rail transit significantly reduces congestion costs. In fact, transit congestion cost savings more than offset total rail transit subsidies. A comprehensive rail transit system can reduce per capita congestion delays by half, and even greater reductions probably occur on specific corridors. However, this does not mean that such cities lack congestion. In fact, congestion, measured as roadway level of service or average traffic speeds, is often quite intense in these cities. However, people in these cities have travel alternatives available on congested corridor, and tend to drive less, and so they experience significantly less congestion delay each year.

Critics sometimes claim that there is no evidence that rail transit reduces traffic congestion, ignoring the evidence presented in this and other studies. In some cases they use analysis which ignores differences in city size, therefore concluding incorrectly that rail transit causes congestion. They often use inappropriate congestion indicators, such as the Travel Time Index, which only measures delay per unit of roadway (automobile and bus) travel, and so ignores delay reductions when people shift to rail, and from more accessible land use patterns that reduce travel distances. This index actually implies that congestion declines if residents increase their vehicle mileage and total travel time, for example, due to more dispersed land use, provided the additional driving occurs in less congested conditions.

Rail transit systems may appear costly due to various special factors:

• New transit projects must overcome decades of underinvestment in grade-separated transit.

• Transit must provide a high quality of service to attract discretionary riders out of their cars.

• Rail transit is generally constructed in the densest part of a city where any transportation project is costly, due to high land values, numerous design constraints, and many impacts.

• Rail transit projects often include special amenities such community redevelopment and streetscape improvements which provide additional benefits, besides just mobility.

• Rail transit projects include tracks, trains, stations, and sometimes parking facilities. It is inappropriate to compare rail system costs with just the cost of adding roadway capacity; comparisons should also include vehicle and parking costs needed for automobile travel.

Table 4 Typical Automobile Commute Trip Costs (Litman, 2003b)

Most people never purchase a road or individual parking space and so greatly underestimate the full cost of accommodating additional urban automobile travel, taking into account vehicle, road and parking costs. Table 4 and Figure 14 show typical estimates of these costs.

Figure 14 Average Operating Costs By Transit Mode (APTA, 2002; Litman, 2003b)

Critics often claims that rail transit is more costly than bus or automobile transport, but this often reflects faulty analysis. They usually consider just a small portion of total transit benefits and underestimate the actual costs of accommodating additional automobile travel under the same conditions, taking into account the high costs of increasing road and parking capacity on major urban corridors. When all benefits and costs are considered, rail transit often turns out to be the most cost effective way of accommodating additional urban travel.

Claims that rail transit projects consume an excessive portion of transportation budgets also tend to reflect incomplete analysis. For example, transportation expenditures by federal, state and local governments totaled $167 billion in 2000, of which $104 billion was for roads, $15.9 billion for bus transit, $1.8 billion for demand response services and $16.7 billion for rail. The cost of parking at destinations is estimated to total more than $200 billion annually (Litman, 2003b). Rail transit expenditures equal about 5% of total automobile facility costs (roads and parking), as illustrated in Figure 15.

Figure 15 Transportation Expenditures (Litman, 2003b; BTS, 2003, Table 3-29a)

When a major rail transit project is under construction most of the cost is included in a particular transportation agency’s capital budget, so for a few years it appears relatively large. This is no different than other major investments, including highway projects and bridges, or a household’s automobile purchase, which may appear exceptionally large compared with a single year’s budget. When averaged over a larger time period (rail transit capital investments have 20-50 year operating lives), or over several cities, transit capital projects represent a small portion of total government transportation expenditures.

Rail systems are sometimes justified for special reasons. For example, New Orleans and Seattle have popular tourist trolley systems which have high costs per passenger-mile, because they are small and serve short trips, but are considered worthwhile investments because they contribute a special ambiance and attract visitors. Rail transit may also be considered worthwhile to support strategic development objectives, or to allow a commercial center to grow. It is simply not economically possible for a center to expand beyond about 5,000 employees without a significant portion of commuters arriving by transit, due to limited road and parking capacity. Because diesel buses are noisy and smelly, large bus terminals are less suitable than rail stations for accommodating large numbers of transit passengers. Although rail systems may seem costly, a significant portion of their costs are often offset by increased property values, business activity and productivity gains (Smith and Gihring, 2003).

Special care is needed when comparing automobile and transit funding. Transit is funded to help achieve various objectives, including congestion reduction, road and parking facility cost savings, consumer cost savings, basic mobility for disadvantaged people, increased safety, pollution reduction and support for strategic development objectives. For efficiency-justified funding (to reduce costs such as congestion, facility costs, accidents and pollution) transit and automobile transport can be compared using measures of cost effectiveness, such as costs per passenger-mile or benefit/cost ratio, to identify the cheapest option. In that case, there is no particular reason to subsidize a transit trip more than an automobile trip, provided all costs (including road and parking costs, traffic services, congestion and crash risk impacts on other road users, and environmental impacts) are considered.

However, for equity-justified service (providing basic mobility to disadvantaged people) there are reasons to subsidize transit more than automobile travel, because transit bears additional costs to accommodate people with disabilities (such as wheelchair lifts), and many non-drivers have low incomes, so greater public subsidies are justified on equity grounds. Since many of these people cannot drive, the alternative must include the cost of a driver, so transit costs should be compared with taxi service costs (or a combination of taxi and chauffeured automobile travel, taking into account the value of time by family members and friends who drive), not simply with vehicle costs.

Care is also needed when comparing different types of transit. Buses are generally cheaper to operate than trains per vehicle-mile, but trains have more capacity and so are cheaper per passenger-mile on routes with high demand. Similarly, costs per vehicle-mile or vehicle-hour tend to be higher in larger cities, due to increased congestion and higher wages, but ridership also tends to be higher, reducing costs per passenger-mile.

Figure 16 Average Operating Cost By Mode and City Category (APTA, 2002)

Operating costs per transit passenger-mile are generally lower in Large Rail cities than in Small Rail cities, and heavy and commuter rail costs are lower than light rail and bus costs, as illustrated in figures 16 and 17.

Figure 17 Operating Cost By Mode And City Category (APTA, 2002)

Large Rail transit systems tend to have lower operating costs than Small Rail systems.

Rail transit systems also tend to have greater cost recovery, that is, a larger portion of operating costs are paid by fares, as illustrated in Figure 18. Transit cost recovery (including both rail and bus services) averages 38% for Large Rail systems (36% excluding New York), 24% for Small Rail systems, and 21% for Bus Only systems.

Figure 18 Transit System Cost Recovery (FTA, 2002)

Some critics argue that rail transit absorbs an excessive portion of transit funding, reducing funding for bus services. But total transit funding tends to increase with rail service as indicated in Figure 19. Thompson and Matoff (2003) find that Bus Only cities such as Columbus, Ohio spend less per capita on transit than cities with rail systems, such as Portland, San Diego and Seattle. This suggests that rail and bus investments are complements rather than substitutes, because decision-makers realize the importance of creating an integrated transit system. This may not be true in every case, but there is no evidence that rail system development necessarily reduces bus funding or service quality.

Figure 19 Annual Per Capita Transit Expenditures

To the degree that transit substitutes for automobile travel, it reduces road and parking facility costs. Table 5 illustrates an estimate of these savings, based on estimates of automobile trip substitution rates, and cost values from Table 4.

Table 5 Estimated Road and Destination Parking Cost Savings

	Large Rail	Small Rail	Totals
Transit Passenger-Miles (millions)	32,107	8,957
Portion of Transit Passenger-Miles by Rail	80%	31%
Portion of transit trips that substitute for a car trip.	60%	50%
Avoided Roadway Costs (cents per veh.-mile)	$0.50	$0.25
Total Roadway Cost Savings (millions)	$7,697	$349	$8,046
Avoided Parking Costs (cents per vehicle-mile)	$0.40	$0.30
Total Parking Cost Savings (millions)	$6,158	$419	$6,577
Total Road and Parking Savings (millions)	$13,855	$768	$14,623

These estimates are conservative because they do not account for the additional savings from the automobile trip reductions leveraged by rail transit, due to reductions in vehicle ownership and improved accessibility due to transit oriented development. Residents in such communities walk rather than drive for more local errands, providing additional road and parking cost savings for those trips.

In addition, reduced vehicle ownership provides residential parking cost savings. Residential parking costs range from about $400 annually for a surface lot in an area with low land values, up to $2,600 annually for underground parking (Litman, 2004a). Parking costs tend to be particularly high in dense urban areas, so it is reasonable to estimate that parking costs average at least $800 in rail transit cities. Rail transit city residents would need to park 6.1 million more vehicles if they owned automobiles at the same rate as Bus Only city residents. At $800 per space, residential parking cost savings for these vehicles total $4.8 billion. Total road and parking cost savings from rail therefore total more than $20 billion dollars annually, substantially more than total rail transit subsidies.

Personal transportation is a major consumer financial burden. About 18% of household expenditures are spent directly on vehicles and transit fares (BLS, 2003). Rail transit provides significant consumer savings. Large Rail city residents spend an average of $2,808 on vehicles and transit, compared with $3,350 in Small Rail cities, and $3,332 in Bus Only cities, despite higher incomes and longer average commute distances. Figures 20 and 21 illustrate these differences.

Figure 20 Transport Expenditures (BLS, 2003)

Large Rail city residents save $22.6 billion in total compared with what consumers spend on transportation in Bus Only cities. These savings are greater than all transit subsidies in the U.S., indicating substantial net economic benefits.

Figure 21 Annual Per Capita Consumer Expenditures on Transportation

Figure 22 Percent Transport Expenditures (BLS, 2003)

Figures 22 and 23 compare transportation as a percentage of household expenditures, which takes into account the higher wages in large cites. Large Rail city residents devote just 12.0% of their income to transportation (this does not change if New York is excluded), compared with 15.8% in Small Rail cities, and 14.9% in Bus Only cities.

Figure 23 Percent Transport Expenditures

Traffic accidents impose significant costs. Despite significant traffic safety efforts, vehicle accidents continue to be the largest cause of deaths and disabilities for people in the prime of life, imposing many billions of dollars in economic losses annually.

Figure 24 Traffic Deaths (FTA, 2001)

Rail transit cities have significantly lower per capita traffic death rates, as illustrated in Figures 24 and 25. Large Rail cities average 7.5 traffic fatalities per 100,000 population (7.9 excluding New York), Small Rail cities average 9.9, and Bus Only cities average 11.7, a 40% higher rate. If Large Rail cities had the same fatality rate as Bus Only cities there would be 251 more annual traffic deaths, plus increased disabilities, injuries and property damages. This represents $5.6 billion in annual savings, based on USDOT recommended values for valuing crash reduction benefits.

Figure 25 Annual Per Capita Traffic Deaths

Rail transit can provide substantial energy conservation and emission reduction benefits. Rail travel consumes about a fifth of the energy per passenger-mile as automobile travel, due to its high mechanical efficiency and load factors (Figure 26). Electric powered rail produce minimal air and noise emissions. Rail provides even greater energy and emission reduction benefits when it leverages additional reductions in vehicle travel.

Figure 26 Transit Energy Consumption (Shapiro, Hassett, and Arnold)

Residents of Large Rail cities drive 12-20% fewer vehicle-miles than residents of Small Rail or Bus Only cities, due to rail’s leverage effect on vehicle ownership and land use. This suggests that rail transit can provide about half the per capita transportation CO₂ emission reductions required to meet the Kyoto targets. In addition:

• Rail transit emission reductions can be particularly large since transit oriented development tends to reduce short automobile trips, in which energy consumption and pollution emissions are high per vehicle mile due to cold starts, and because these trips occur under congested conditions. As a result, each 1% of mileage reduced typically reduces air emissions by 2-3%.

• Rail tends to reduce emissions in highly populated areas, such as city centers, major roadways and transit terminals, and so reduces people’s exposure to harmful emissions such as CO, toxics and particulates, particularly compared with diesel buses.

• Transit encouragement strategies that increase ridership, and transit oriented development policies, tend to have large energy conservation and emission reduction benefits.

• Energy conservation and pollution emission reductions are just two of many potential benefits of rail transit. When these additional benefits are considered, rail investments can be a cost effective way to achieve environmental objectives.

Economic Development refers to progress toward a community’s economic goals, including increased productivity, employment, business activity, investment and redevelopment. Transit in general and rail transit in particular can provide a variety of economic development benefits (Cambridge Systematics, 1998; Forkenbrock and Weisbrod, 2001; MKI, 2003; Litman, 2004a). These benefits are summarized below.

As described earlier, by attracting discretionary travelers, increasing transit ridership, and providing a catalyst for more efficient land use, rail transit provides various cost savings and efficiency gains, including congestion reduction, road and parking cost savings, consumer savings, reduced crash damages, and improved public health. These economic savings and efficiency benefits filter through the economy as savings to consumers, businesses and governments, making a region more productive and competitive.

Expenditures on automobiles, fuel and roadway facilities provide relatively little regional economic activity because they are capital intensive and largely imported from other areas. A study using national input-output table data found that each 1% of regional travel shifted from automobile to public transit increases regional income about $2.9 million, resulting in 226 additional regional jobs (Miller, Robison & Lahr, 1999). These impacts are summarized in Table 6. As described earlier, Large Rail city residents spend an average of $448 less annually per capita on transportation than residents of Bus Only cities, despite higher incomes and longer average commute distances, totaling $22.6 billion in savings. If each million dollars in consumer expenditures shifted from automobile expenses to general consumer expenditures provides an average of 8.6 jobs and $219,000 in regional income, as indicated in Table 6, rail transit provides a total of 194,114 additional jobs and $4.9 billion in additional regional income in those cities.

Table 6 Regional Economic Impacts of $1 Million Expenditure

Expenditure Category	Regional Income	Regional Jobs
Automobile Expenditures	$307,000	8.4
Non-automotive Consumer Expenditures	$526,000	17.0
Transit Expenditures	$1,200,000	62.2

Land use density and clustering tend to provide agglomeration benefits, which can reduce the costs of providing public services and increase productivity due to improved accessibility and network effects (Litman, 2003c). One published study found that doubling a county-level density index is associated with a 6% increase in state-level productivity (Haughwout, 2000). This suggests that transit improvements can help create land use patterns that increase regional productivity and economic development. Although these impacts are difficult to measure, they are likely to be large.

Transit oriented development tends to increase local property values due to improved accessibility and livability in that area (Eppli and Tu, 2000; Smith and Gihring, 2003). Transit stations often provide a catalyst for various neighborhood improvements such as urban redevelopment, historic preservation, improved pedestrian conditions and New Urbanist design practices. A portion of these property value gains may be economic transfers (property value increases in one area are offset by property value reductions at other locations), but increased property values resulting from agglomeration efficiencies, shifted consumer expenditures, transportation efficiency and community redevelopment are true economic gains that increase productivity.

Current development patterns tend to abandon older neighborhoods as new communities are built at the urban fringe. This tends to be inefficient in terms of infrastructure (roads, schools and other facilities in urban areas are underused while new facilities must be built in suburban areas) and in terms of social capital (many older neighborhoods have unique cultures, traditions and human relationships). This results, in part, from growing automobile traffic through older neighborhoods caused by urban fringe residents. Rail transit can provide a catalyst for urban redevelopment and help reduce automobile traffic volumes through urban areas. A unique transit service can be a popular tourist activity, help create community identity, which stimulates economic development.

Transit in general, and rail transit in particular, can provide important but difficult to measure benefits (Forkenbrock and Weisbrod, 2001). These are described briefly below.

Automobile-dependent transport and land use patterns disadvantages non-drivers. It also imposes costs on motorists, who are forced to chauffeur non-driving family members and friends. Transit improvements and transit oriented development increase mobility and accessibility options for non-drivers. Since non-drivers tend to be physically, economically and socially disadvantaged compared with drivers, this increases equity, in addition to reducing costs and increasing economic productivity.

Chauffeuring refers to additional automobile travel specifically to carry a passenger. It excludes ridesharing, which means additional passengers in a vehicle that would be making a trip anyway. Some motorists spend a significant amount of time chauffeuring children to school and sports activities, family members to jobs, and elderly relatives on errands. Such trips can be particularly inefficient if they require drivers to make an empty return trip, so a five-mile passenger trip produces ten miles of total vehicle travel. Drivers sometimes enjoy chauffeuring, for example, when it gives busy family members or friends time to visit. However, chauffeuring can be an undesirable burden, for example, when it conflict with other important activities. Quality transit service and transit oriented development allows drivers to avoid undesirable chauffeuring trips.

Transit services provide option value, referring to the value people place on having a service available even if they do not currently use it (ECONorthwest and PBQD, 2002). Transit provides critical transportation services during personal and community-wide emergencies, such as when a personal vehicle has a mechanical failure, or a disaster limits automobile travel.

Community Livability refers to the environmental and social quality of an area as perceived by residents, employees, customers and visitors. Rail transit and transit oriented development can help improve community livability in several ways, including urban redevelopment, reduced vehicle traffic, reduced air and noise pollution, improved pedestrian facilities, and greater flexibility in parking requirements and street design. This provides direct benefits to residents, increases property values and can increase retail and tourist activity in an area.

Many people lead overly-sedentary lifestyles, which causes various health problems. Increased walking is one of the most popular and effective way to increase physical activity among otherwise sedentary people. To the degree that transit trips involve walking or cycling links, and transit oriented development improves walking and cycling conditions, it can improve public health.

Table 7 summarizes U.S. transit service expenditures and revenues. Rail subsidies (operating and capital expenses minus fare revenues) totaled $12.5 billion in 2002, averaging about $140 per capita when divided among the 90 million residents of cities with rail transit systems, compared with $13.8 billion bus transit subsides, which averages about $50 per capita when divided among 278 million U.S. residents. This indicates that the incremental cost of rail transit is about $90 annually per capita.

Table 7 U.S. Transit Expenses and Revenues By Mode (APTA, 2002)

This compares with $67.7 billon in estimated monetized (measuring in monetary units) benefits identified in this study, as summarized in Table 8. This indicates that, considering just impacts suitable for monetization, economic benefits are many times greater than subsidy costs. Rail transit provides additional benefits that are unsuited to monetization, including economic development, improved mobility for non-drivers, community livability and improved public health. Even people who do not currently use rail transit benefit from reduced traffic and parking congestion, and other benefits that disperse through the economy.

Table 8 Rail Transit Monetized Benefits

There is considerable debate over the relative merits of bus and rail transit (Pascall, 2001; GAO, 2001; Warren and Ryan, 2001; Thompson and Matoff, 2003; Balaker, 2004). Some key issues are discussed here.

Rail transit tends to provide better service quality that attracts more riders, particularly discretionary users. Rail can carry more passengers per vehicle which reduces labor costs, requires less land per peak passenger-trip, and causes less noise and air pollution compared with diesel buses. As a result, rail is more suitable for high-density areas. Voters are often more willing to support funding for rail than for bus service. Transit-oriented land use patterns can increase property values and economic productivity by improving accessibility, reducing costs, improving livability and providing economies of agglomeration. In some cases, increased property values offset most or all transit subsidy costs. This does not generally occur with bus service.

Rail transit can be compared to a luxury vehicle: it costs more initially but provides higher quality service and greater long-run value. As consumers become wealthier and accustomed to higher quality goods it is reasonable that they should demand features such as more leg-room, comfortable seats, smoother and quieter ride (and therefore better ability to read, converse, and rest), and greater travel speed associated with grade-separated transit. The preference of rail over bus can be considered an expression of consumer sovereignty, that is, people’s willingness to pay extra for the amenities they prefer.

Bus transit tends to be cheaper to develop and more flexible. Proponents argue that bus service can be as fast and comfortable as rail, and that much of the preference for rail reflects prejudices rather than real advantages. Bus transit can serve a greater area, and so can attract greater total ridership than rail with comparable resources, particularly in areas with dispersed destinations. Buses tend to provide basic mobility services used by people who are transportation disadvantaged, and so tends to provide greater equity benefits.

Key differences between bus and rail transit are summarized on the next page. Rather than a debate about which is better, each can be considered most appropriate in particular situations. Bus is best serving areas with more dispersed destinations and lower demand. Rail is best serving corridors where destinations are concentrated, such as large commercial centers and mixed-use urban villages. Rail can be a catalyst for creating more accessible, multi-modal communities and urban redevelopment. Rail tends to attract more riders within a given area, but buses can cover more area. Both can become more efficient and effective at achieving planning objectives if implemented with supportive policies that improve service quality, create supportive land use patterns and encourage ridership.

Rail transit is not appropriate in every situation, and even the most best transit program can still be improved. Rail transit supporters should therefore welcome legitimate criticism to help identify possible problems and opportunities for improvement. However, some types of criticism are not helpful, because they misrepresent issues and reflect inaccurate analysis. It is therefore helpful to examine and evaluate rail transit criticisms to identify legitimate issues and concerns, and to recognize errors and misrepresentations.

A good research document provides readers with the information they need to make an informed assessment, including an overview of issues and information sources, discussion of various perspectives and evaluation methods, and information that both supports and contradicts (if any exists) the authors conclusions (Litman, 2004b). Many transit studies do this, providing accurate and useful analysis.

But some critics provide inaccurate information and biased analysis intended to present rail transit in a negative. They fail to use best practices for accurate transit evaluation. They ignoring other perspectives, and suppress data that contradict their arguments. These critics tend to consider a relatively limited set of transit impacts, as summarized in Table 9. As a result, they tend to understate the full benefits of transit.

Table 9 Impacts Considered and Overlooked (Litman, 2004a)

Specific examples of rail transit criticism are examined below.

"Great Rail Disasters" (O’Toole, 2004)

Great Rail Disasters argues that rail transit is ineffective at improving transportation system performance and wasteful. Other rail critics, such as Balaker (2004), have citied O’Toole’s study heavily. Great Rail Disasters uses a thirteen component index created by the author to evaluate rail transit system performance. This analysis framework appears to be carefully designed to portray rail transit in a negative way. The report contains several fundamental omissions and misrepresentations. Major errors include:

• Failing to differentiate between cities with relatively large, well-established rail systems and those with smaller and newer systems that cannot be expected to have significant impacts on regional transportation performance.

• Lack of with-and-without analysis. There are virtually no comparisons between cities that have rail and those that do not. It is therefore impossible to identify rail transit impacts.

• Evaluating congestion impacts based on "Travel Time Index" values. Of the various congestion indicators this is one of the least appropriate for evaluating grade-separated transit, since it only considers delays to road vehicles, ignoring benefits to people who shift to transit, and from vehicle traffic reductions due to more accessible land use.

• Failing to compare individual city’s and national trends. During the time period used for analysis, from 1970 to 2000, transit ridership and mode split declined nationally, so a lower rate of decline could be considered successful compared with most other cities.

• Failing to account for additional factors that affect transportation and urban development conditions, such as city size, changes in population and employment.

• Ignoring and understating significant costs of automobile travel. Vehicle expenses are included when calculating transit costs, but vehicle and parking expenses are ignored when calculating automobile costs.

• Exaggerating transit development costs. Claims, such as "Regions that emphasize rail transit typically spend 30 to 80 percent of their transportation capital budgets on transit" are unverified and generally only true for certain regions and years, not when costs are averaged over larger areas and times.

• Presenting outdated data as current, including examples from the 1960s through early 80’s, and airport ridership data from 1990.

• Ignoring other benefits of rail transit, such as parking cost savings, consumer cost savings and increased property values in areas with rail transit systems.

• Failing to reference documents that reflect current best practices in transit evaluation, such as ECONorthwest and PBQD (2002) or Litman (2004) or provide any information showing alternative perspectives.

Great Rail Disasters’ bias is revealed in its analysis of Portland, Oregon. According to many of its own indicators Portland’s rail system is successful, with increasing transit ridership and commute mode split. Still, O’Toole concludes that Portland’s rail system is harmful because it involves transit oriented development, which he opposes on the grounds that it is harmful to consumers. Yet, there is plenty of evidence that many consumers want to live in transit oriented communities.

"Light Rail Boon or Boondoggle" (Castelazo and Garrett, 2004)

An article by Molly D. Castelazo and Thomas A. Garrett ("Light Rail: Boon or Boondoggle" 2004) argues that light rail investments are inefficient. Their analysis contains several critical errors. They ignore many costs of automobile transportation, including roadway costs, consumer costs, downstream congestion, parking facility costs, accident costs and pollution impacts. They use average cost values that underestimate the actual costs of accommodating increased automobile traffic in dense urban areas. They claim that light rail is more costly than automobile or bus transport, based on a national cost value of 54.4¢ per passenger-mile for light rail, although the actual cost in St. Louis is just 27¢, which is lower than either automobile or bus costs. They claim that light rail only provides short-term congestion and pollution reduction benefits, which is untrue, and indicates that they are unfamiliar with the issues.

Castelazo and Garrett argue that it would be cheaper to provide low-income motorists with a car than light rail transit service. This overlooks several important points.

• First, transit is subsidized several reasons besides providing mobility to lower-income travelers. Only a small portion of transit subsidies could efficiently or equitably be shifted to any one of these objectives.

• Second, many transit riders cannot or should not drive. Subsidized cars would not solve their mobility problems, and would tend to increase higher-risk driving.

• Third, substituting car ownership for transit service is more expensive than they claim. Eliminating scheduled transit service would force riders who cannot drive to use demand-response or taxi services, which have far higher costs than simply driving a car.

• Fourth, increased vehicle traffic on busy urban corridors would significantly increase traffic congestion, road and parking costs, accidents, pollution and other external costs. Castelazo and Garrett underestimate these costs. In footnote 3 they calculate that giving 7,700 vehicles to current rail users would only increase regional congestion by 0.5%. But rail users commute on the city’s most congested corridors, so congestion impacts will be proportionately large. The Texas Transportation Institute calculates that St. Louis traffic congestion costs totaled $738 million in 2001. If 7,700 additional downtown automobile commuters increases congestion 2.5-5.0%, this represents $18 to $37 million in additional annual congestion costs.

• Fifth, there are substantial practical problems subsiding cars. Castelazo and Garrett apparently assume that the 7,700 rail transit riders they identify as being unable to afford a car are a distinct, identifiable group. In fact, they consist of a much larger group, many of whom only use transit occasionally. As a result, it would be necessary to offer a much larger number of households a part-time car, with provisions that account for constant changes in their mobility needs and abilities. Like any subsidy program, it would face substantial administrative costs and require complex rules to determine who receives how much subsidy in a fair and effective way. It would create perverse incentives, rewarding poverty and automobile dependency.

• Finally, as described earlier, rail transit can provide a catalyst for mixed-use, walkable urban villages and residential neighborhoods where it is possible to live and participate in normal activities without needing an car, which is particularly beneficial to non-drivers.

"Urban Rail: Uses and Misuses" (Cox, 2000)

Wendell Cox is a frequent critic of rail transit. He makes the following claims in a policy statement titled Urban Rail: Uses and Misuses. Responses to his claims are in italics.

• Virtually no traffic congestion reduction has occurred as a result of building new urban rail systems.